Naming of trend calculation files
*********************************

Each text file name has multiple pieces that indicate what trend analysis runs 
are represented in it.

Quantity - freq, alb, iwc = occurrence frequency, albedo, ice water content

Hemisphere - nh, sh = North, South

Date Range - m20p55 = [-20,+55] DSS 

Local Time Adjustment - raw, ltconst, ltflat = no adjustment, "constant" 
    adjustment, "flat" adjustment as discussed in report

Fit Range - trend, split = single segment, two segments for linear term

The only exception is that for the two-segment calculations, the raw and 
constant local time results are combined without having a tag in the file name 
(e.g. iwc_nh_m20p55_split.txt).


Contents of trend calculation files
***********************************

Within each "single segment" trend output file, there are three blocks of data 
for the different normalization methods.  Each block has 4 lines for the 
different latitude bands that contain the following information:
  - latitude band
  - "best fit" phase lag (minimum chi-square)
  - fit coefficients with 1 sigma values
  - average value over the period (used to calculate percent changes)
  - individual correlation coefficients and multiple correlation
  - solar cycle amplitude [%] (peak-to-peak range = 2.5e+11 phot/cm^2/sec/nm)
  - long-term trend [percent/decade]
  - RMS scatter in data units

Below that are 4 more lines that give the 95% significance level for decadal 
trend as calculated by the Weatherhead et al. formula given in our 2007 paper.

NOTE:  Some files show '***' for the IWC solar coefficient at 74-82 degrees 
because I didn't properly adjust the print format statement.  That error has 
been corrected.